Abstract

Optimising data collection work has always been a key challenge in conducting statistical surveys. When using traditional paper questionnaires, it is difficult to implement logical and consistency checks on the collected data, and the approach will inevitably requires subsequent data input that is time consuming and error-prone. Introducing multi-modal data collection, vis. online questionnaire (OQ), telephone interview (TI) and face-to-face interview (FI), typically leads to the collected data scattered in different paper and digital formats if relevant workflows and systems are not integrated tightly. Complicated manual data processing and consolidation efforts become necessary. Meanwhile, management of different fieldwork activities may be done separately for individual surveys, making it difficult to achieve overall efficiency. Furthermore, the existence of a number of different survey-specific computer systems involves high maintenance cost, and the approach is not conducive to meeting fast changing data needs.

To streamline the data collection procedures of various statistical surveys, facilitate efficient and effective data quality controls, and enhance fieldwork management, the Census and Statistics Department (C&SD) of Hong Kong, China is developing Data Collection Platform (DCP) and Fieldwork Management System (FMS) as centralised and scalable setups to meet the aforementioned objectives.

This new DCP and FMS are purposefully designed to support multiple surveys conducted by C&SD, regardless of household or establishment types. At its frontend, DCP has adopted progressive web application design that can support OQ for self-reporting, TI on workstations and FI on mobile tablets at the same time. At the backend, DCP can transform, synchronise, and store data gathered through different modes over the same data infrastructure and thus will significantly reduce the time and resource spent on subsequent data processing stage.

FMS is closely integrated with DCP to provide essential fieldwork management functions, including assignment and re-assignment of caseloads, itinerary planning, quality controls, progress monitoring and performance management.

The development of DCP and FMS is as much a business process re-engineering exercise as it is a computer system development project. Relevant workflows are streamlined and standardised across different surveys to maximise commonality, as far as possible, for building centralised platforms. User groups are formed not only to tap users’ experience and knowledge, but also to get their early engagement to prepare for the changes. This is crucial to ensure the project can progress swiftly within tight schedule. This paper introduces the key features of the new centralised DCP and FMS and the perceived benefits in greater details using the General Household Survey as an example.